Effects of drugs are in clinical trials often measured on categorical scales. These measurements are increasingly being analyzed using mixed-effects logistic regression. However, the experience with such analyzes is limited and only a few models are used. The aim of this thesis was to investigate the performance and improve the use of models and methods for mixed-effects categorical data analysis. The Laplacian method was shown to produce biased parameter estimates if (i) the data variability is large or (ii) the distribution of the responses is skewed. Two solutions are suggested; the Gaussian quadrature method and the back-step method. Two assumptions made with the proportional odds model have also been investigated. The assumption with proportional odds for all categories was shown to be unsuitable for analysis of data arising from a ranking scale of effects with several underlying causes. An alternative model, the differential odds model, was developed and shown to be an improvement, in regard to statistical significance as well as predictive performance, over the proportional odds model for such data. The appropriateness of the likelihood ratio test was investigated for an analysis where dependence between observations is ignored, i.e. performing the analysis using the proportional odds model. The type I error was found to be affected; thus assessing the actual critical value is prudent in order to verify the statistical significance level. An alternative approach is to use a Markov model, in which dependence between observations is incorporated. In the case of polychotomous data such model may involve considerable complexity and thus, a strategy for the reduction of the time-consuming model building with the Markov model and sleep data is presented. This thesis will hopefully contribute to a more confident use of models for categorical data analysis within the area of pharmacokinetic and pharmacodynamic modelling in the future.

Exposure to antibiotics is an important factor influencing the development of bacterial resistance.  In an era where very few new antibiotics are being developed, a strategy for the development of optimal dosing regimen and combination treatment that reduces the rate of resistance development and overcome existing resistance is of utmost importance. In addition, the optimal dosing in subpopulations is often not fully elucidated. The aim of this thesis was to develop pharmacokinetic (PK) and pharmacokinetic-pharmacodynamic (PKPD) models that characterize the interaction of antibiotics with bacterial growth, killing and resistance over time, and can be applied to guide optimization of dosing regimens that enhance the efficacy of mono- and combination antibiotic therapy. A mechanism-based PKPD model that incorporates the growth, killing kinetics and adaptive resistance development in Escherichia coli against gentamicin was developed based on  in vitro time-kill curve data. After some adaptations, the model was successfully applied for similar data on colistin and meropenem alone, and in combination, on one wild type and one meropenem-resistant strain of Pseudomonas aeruginosa. The developed population PK model for colistin and its prodrug colistin methanesulfonate (CMS) in combination with the PKPD model showed the benefits for applying a loading dose for this drug. Simulations predicted the variability in bacteria kill to be larger between dosing occasions than between patients. A flat-fixed loading dose followed by an 8 or 12 hourly maintenance dose with infusion duration of up to 2 hours was shown to result in satisfactory bacterial kill under these conditions. Pharmacometric models that characterize the time-course of drug concentrations, bacterial growth, antibacterial killing and resistance development were successfully developed. Predictions illustrated how PKPD models based on in vitro data can be utilized to guide development of antibiotic dosing, with examples advocating regimens that (i) promote bacterial killing and reduce risk for toxicity in preterm and term newborn infants receiving gentamicin, (ii) achieve a fast initial bacterial killing and reduced resistance development of colistin in critically ill patients by application of a loading dose, and (iii) overcome existing meropenem resistance by combining colistin and meropenem

Current excessive use and abuse of antibiotics has resulted in increasing bacterial resistance to common treatment options which is threatening to deprive us of a pillar of modern medicine. In this work methods to optimize the use of existing antibiotics and to help development of new antibiotics were developed and applied. Semi-mechanistic pharmacokinetic-pharmacodynamic (PKPD) models were developed to describe the time course of the dynamic effect and interaction of combinations of antibiotics. The models were applied to illustrate that colistin combined with a high dose of meropenem may overcome meropenem-resistant P. aeruginosa infections. The results from an in vivo dose finding study of meropenem was successfully predicted by the meropenem PKPD model in combination with a murine PK model, which supports model based dosage selection. However, the traditional PK/PD index based dose selection was predicted to have poor extrapolation properties from pre-clinical to clinical settings, and across patient populations. The precision of the model parameters, and hence the model predictions, is dependent on the experimental design. A limited study design is dictated by cost and, for in vivo studies, ethical reasons. In this work optimal design (OD) was demonstrated to be able to reduce the experimental effort in time-kill curve experiments and was utilized to suggest the experimental design for identification and estimation of an interaction between antibiotics. OD methods to handle inter occasion variability (IOV) in optimization of individual PK parameter estimates were proposed. The strategy was applied in the design of a sparse sampling schedule that aim to estimate individual exposures of colistin in a multi-centre clinical study. Plasma concentration samples from the first 100 patients have been analysed and indicate that the performance of the design is close to the predicted. The methods described in this thesis holds promise to facilitate the development of new antibiotics and to improve the use of existing antibiotics.

Pharmacodynamic studies are important for the optimal use of antimicrobial agents. Combination antifungal therapy may be one method to improve outcome in invasive Candida infections. An in vitro kinetic model to study the pharmacodynamic effects of a combination of two antifungal agents with different elimination rates was developed and the pharmacodynamics of amphotericin B (AMB), voriconazole (VRC) or the combination was evaluated. Exposure to VRC inhibited the fungicidal activity of sequential doses of AMB against VRC-susceptible strains of C. albicans. The interaction was VRC dose-dependent. AMB activity was regained once VRC was removed or it increased gradually when the concentration of VRC had fallen below the minimum inhibitory concentration (MIC). The VRC-AMB interaction, however, was also present against strains of C. albicans, C. glabrata and C. krusei despite reduced VRC susceptibility. Against these strains the interaction was not predicted by the MIC value, suggesting that mechanisms of resistance may be of importance. Until more data are available, a reasonable recommendation is probably to avoid the sequential use of VRC followed by AMB and to use the combination of VRC and AMB for the treatment of Candida infections with caution. Only the unbound fraction of a drug is generally accepted as pharmacologically active. The activity of posaconazole (POS) with a protein binding of 98-99% was tested in serum against Candida species and compared with the calculated unbound serum concentration in protein-free media. Significant differences emerged at clinically relevant POS serum concentrations of 1.0 and 0.10 mg/l compared with the serum control regimen against one strain of C. lusitaniae. In RPMI 1640 the corresponding calculated unbound concentrations resulted in no effect for the low dose regimen compared with the RPMI 1640 control regimen. Further, against seven additional Candida strains tested, the effect of POS was greater in serum than in RPMI 1640. A flux from serum protein bound to fungal lanosterol 14α-demethylase bound POS may be the explanatory mechanism.

In critically ill patients, the pathophysiology of sepsis can affect the interactions between the antibiotic, the bacteria and the patient, leading to potential therapeutic failure and the development of antibiotic resistance. It is well acknowledged that research that optimises antibiotic exposure will assist improvement of outcomes in this patient group. Although beta-lactam antibiotics, such as piperacillin and meropenem, are commonly selected for empiric therapy of sepsis, dosing is unlikely to be optimal. In patients without renal dysfunction, data suggests that disease-induced alterations to pharmacokinetic parameters result in low trough concentrations for significant periods within a dosing interval. Administration of these time-dependent antibiotics by continuous infusion has been suggested to improve the pharmacokinetic-pharmacodynamic profile. Knowledge of concentrations in the extracellular fluid of human tissue, which is the target site of most pathogens, is particularly instructive. Extracellular fluid concentrations can be determined using techniques such as microdialysis. Therefore, the principal aims of this Thesis were to determine the plasma and subcutaneous tissue pharmacokinetics of piperacillin and meropenem administered by bolus dosing and continuous infusion in critically ill patients with sepsis; and to use Monte Carlo simulations to compare the ability of different dosing strategies to achieve pharmacodynamic endpoints. This Thesis also sought to compare the clinical outcomes of bolus dosing and continuous infusion of a beta-lactam antibiotic, ceftriaxone, in a prospective randomised controlled trial and to perform a meta-analysis on clinical outcomes from other similar published studies. Finally, this Thesis aimed to systematically review the published literature to determine any correlation between antibiotic dosing and the development of antibiotic resistance. The results of the pharmacokinetic studies, using piperacillin and meropenem, indicate that beta-lactam distribution into subcutaneous tissue, in critically ill patients with sepsis, is less than that observed in previous studies in healthy volunteers yet superior to studies in patients with septic shock. This supports the notion that the peripheral concentration of drugs may be inversely related to the level of sickness severity. Administration by continuous infusion was found to maintain statistically significantly higher trough beta-lactam concentrations in both plasma and subcutaneous tissue. Further analysis of the plasma data using population pharmacokinetic modeling and Monte Carlo simulations described significant pharmacodynamic advantages for administering meropenem or piperacillin by continuous infusion to organisms with high minimum inhibitory concentrations. Given the documented pharmacodynamic advantages for administering beta-lactams by continuous infusion, a prospective randomized controlled clinical trial was conducted using the beta-lactam antibiotic ceftriaxone. In 57 critically ill patients, we found equivalence between continuous infusion and bolus dosing in the intention-to-treat analysis. However, our a priori analysis criteria, requiring patients receive at least 4-days antibiotic treatment, found significant clinical and bacteriological advantages for administration by continuous infusion. To further investigate any clinical differences between bolus dosing and continuous infusion of beta-lactam antibiotics, we performed a meta-analysis of all published studies. Our analysis of the 13 published prospective randomized controlled trials (846 hospitalised patients) showed equivalence of continuous infusion and bolus dosing. Possible confounders observed within, and between the studies, make interpretation of these results challenging. However, two large retrospective cohorts not included in the meta-analysis, found definitive clinical and bacteriological advantages suggesting further research may be appropriate. The possible relationship between antibiotic dosing, or exposures, on the development of resistance was investigated using a structured review of the published literature. Our analysis of relevant papers found a wealth of data describing increasing levels of resistance with sub-optimal antibiotic dosing, particularly for fluoroquinolone antibiotics, but also for other classes including beta-lactams. These results demonstrate the importance of optimizing antibiotic dosing to decrease the development of antibiotic susceptibility from sub-optimal dosing, particularly for critically ill patients who are likely to have low drug concentrations. The results of this Thesis, suggest that a large, prospective, multi-centre randomised controlled trial in critically ill patients with sepsis, is required to definitively determine the clinical utility of administration of beta-lactam antibiotics by continuous infusion.

Φυσιολογικά επίπεδα της αγρίου τύπου απολιποπρωτεϊνης Ε (apoE) στο πλάσμα διαμεσολαβούν στην κάθαρση των αθηρογενετικών λιποπρωτεϊνών ενώ υψηλότερα επίπεδα από τα φυσιολογικά προκαλούν υπερτριγλυκεριδαιμία. Αυτή η ιδιότητα της αγρίου τύπου apoE μειώνει σημαντικά την θεραπευτική της αξία ως ένα πιθανό βιολογικό φάρμακο για την αντιμετώπιση της δυσλιπιδαιμίας. Πρόσφατα, έχει δημιουργηθεί και μελετηθεί μια μεταλλαγμένη μορφή της apoE, apoE4 [ L261A, W264A, F265A, L268A, V269A ] (apoE4mut1) με βελτιωμένες βιολογικές ιδιότητες. Συγκεκριμένα, αυτή η μεταλλαγμένη μορφή μπορεί να φέρει τα υψηλά επίπεδα χοληστερόλης σε φυσιολογικές τιμές χωρίς να προκαλέσει υπερτριγλυκεριδαιμία ακόμα και όταν υπερεκφράζεται. Στην παρούσα μελέτη, πραγματοποιήθηκε φαρμακοδυναμική και φαρμακοκινητική ανάλυση της apoE4mut1 σε πειραματόζωα. Με γονιδιακή μεταφορά μέσω ιού σε ποντίκια που είχαν έλλειψη στον LDL υποδοχέα (LDLr-/-) και σε ποντίκια που είχαν έλλειψη στην apoE (apoE-/-), δείχθηκε οτι η δράση της apoE4mut1 ( μείωση της χοληστερόλης ) εξαρτάται από την έκφραση ενός λειτουργικού κλασσικού LDL υποδοχέα. Εφάπαξ έγχυση της apoE4mut1 συνδεδεμένης με λιποσώματα σε apoE-/- ποντίκια που ήταν σε δίαιτα δυτικού τύπου για 6 εβδομάδες αποκάλυψε οτι η εξωγενώς συντιθέμενη apoE4mut1 διατηρεί άθικτη την ικανότητά της να κανονικοποιεί τα υψηλά επίπεδα χοληστερόλης αυτών των ποντικιών με μια μέγιστη φαρμακολογική απόκριση που παρατηρείται σε μόλις 10 ώρες μετά την έγχυση. Ενδιαφέρον παρουσίασε το γεγονός οτι τα επίπεδα χοληστερόλης του πλάσματος παρέμειναν σημαντικώς μειωμένα για τις επόμενες 24 ώρες μετά την έγχυση της apoE4mut1- λιποσώματα. Μετρήσεις συγκεντρώσεων της apoE έδειξαν οτι η apoE4mut1 στην μορφή των πρωτεολιποσωμάτων που χρησιμοποιήθηκαν σε αυτή τη μελέτη έχει χρόνο ημίσειας ζωής 15.8 h. Τα δεδομένα αυτά οδηγούν στο συμπέρασμα οτι η καθαρή apoE4mut1 μπορεί να αποτελέσει ένα νέο υποψήφιο φάρμακο για την άμεση αντιμετώπιση της υπερχοληστερολαιμίας σε άτομα που εκφράζουν έναν λειτουργικό LDL υποδοχέα. / Physiological levels of wild-type (wt) apolipoprotein E (apoE) in plasma mediate the clearance of cholesterol-rich atherogenic lipoprotein remnants while higher than normal plasma apoE concentrations fail to do so and trigger hypertriglyceridemia. This property of wt apoE reduces significantly its therapeutic value as a potential biological drug for dyslipidemia. Recently, we reported the generation of a recombinant apoE variant, apoE4 [L261A, W264A, F265A, L268A, V269A] (apoE4mut1) with improved biological functions. Specifically, this variant can normalize high plasma cholesterol levels without triggering hypertriglyceridemia, even at supraphysiological levels of expression. In the present study we performed pharmacodynamic and pharmacokinetic analysis of apoE4mut1 in experimental mice. Using adenovirus-mediated gene transfer in LDL receptor deficient (LDLr-/-) and apoE deficient (apoE-/-) mice, we show that the cholesterol lowering potential of apoE4mut1 is dependent on the expression of a functional classical LDLr. Bolus infusion of apoE4mut1-containing proteoliposomes in apoE-/- mice fed western-type diet for 6 weeks indicated that exogenously synthesized apoE4mut1 maintains intact its ability to normalize the high cholesterol levels of these mice with a maximum pharmacological effect obtained at only 10 hours post-treatment. Interestingly, plasma cholesterol levels remained significantly reduced even 24 hours following intravenous infusion of apoE4mut1 proteoliposomes. Measurements of plasma apoE levels indicated that apoE4mut1 in the form of proteoliposomes used in the study has a half-life of 15.8 h. Our data suggest that purified apoE4mut1 may be an attractive new candidate for the acute correction of hypercholesterolemia in subjects expressing functional LDL receptor.

No description available.

La mise en évidence de la variabilité pharmacocinétique et/ou pharmacodynamique permet l’optimisation de l’utilisation des cytotoxiques. L’association des thérapies ciblées à la chimiothérapie conventionnelle peut apporter des avantages supplémentaires en termes de bénéfice thérapeutique, mais aussi peut provoquer des interactions médicamenteuses et augmenter les variabilités interindividuelles. Les interactions médicamenteuses sont maintenant connues pour avoir un impact significatif sur l'élimination des médicaments.Le premier volet de ces travaux consiste à mesurer les concentrations sanguines de deux molécules (le lapatinib et la vinorelbine) chez les patientes et à les utiliser pour la modélisation pharmacocinétique. Cette modélisation a clairement montré l’influence du poids et du nombre des plaquettes sur la pharmacocinétique de la vinorelbine permettant ainsi de diminuer les variabilités pharmacocinétiques. Elle a également mis en évidence l’influence du lapatinib sur la pharmacocinétique de la vinorelbine. Cependant, compte tenu de l’absence de groupe témoin, nous n’avons pas réussi à obtenir une significativité statistique pour ces résultats. En parallèle, aucune influence de la vinorelbine sur le comportement pharmacocinétique du lapatinib n’a été mise en évidence.Le second volet concerne la modélisation de la réponse et de la tolérance des patientes pour cette association médicamenteuse (modélisation pharmacodynamique). La neutropénie est la toxicité dose-limitante et comme le lapatinib n’est pas connu pour être myélotoxique, nous avons modélisé cette toxicité par rapport au comportement pharmacocinétique de la vinorelbine. Là encore, nous avons observé une très forte tendance quant à l’influence du lapatinib sur la myélotoxicité de la vinorelbine. Concernant l’efficacité de cette association, la modélisation de l’évolution de la croissance tumorale a mis en évidence une synergie d’action entre ces deux molécules. A notre connaissance c’est la première fois qu’une telle modélisation tant du point de vue de la tolérance que de celui de la réponse, est réalisée lors d’une association de chimiothérapie classique et d’une thérapeutique ciblée.La modélisation PK-PD de population des médicaments anticancéreux peut apporter une aide précieuse aux cliniciens dans ce domaine. Elle peut également être essentielle dans la prise de décision clinique précoce. / The identification of sources of pharmacokinetic and / or pharmacodynamic variabilities, optimizes the use of cytotoxic agents. The combination of targeted therapies with conventional drugs may provide additional benefits in terms of therapeutic benefit, but also can cause drug interactions and increased variability. Drug interactions are known to have a significant impact on drug disposition.The first part of this work is to measure blood concentrations of two molecules (lapatinib and vinorelbine) in patients and to perform pharmacokinetic modeling. This modeling clearly showed the influence of weight and the platelet number on the pharmacokinetics of vinorelbine reducing thereby the pharmacokinetic variabilities. It also highlighted the influence of lapatinib on the pharmacokinetics of vinorelbine. However, due to the lack of control group, we failed to obtain statistical significance for these results. In parallel, no effect of vinorelbine on the pharmacokinetics of lapatinib has been detected.The second part concerns the modeling of the response and tolerance of patients for this drug combination (pharmacodynamic modeling). Neutropenia was detected as the dose-limiting toxicity of the combination and the fact that lapatinib is not known to be myelotoxic, we performed toxicity modeling based on the pharmacokinetics of vinorelbine. Again, we observed a strong tendency on the influence of lapatinib on the myelotoxicity of vinorelbine. Regarding the efficacy of this combination, tumor growth modeling showed a synergistic action between the two molecules. To our knowledge this is the first time that such a model, both in terms of tolerance and response, is implemented in a combination of conventional chemotherapy and targeted therapy.Population PK-PD modeling of cancer drugs can provide valuable assistance to clinicians in this field. It can also be essential in the early clinical decision making.

Malgré l'existence d'un vaccin et de différentes molécules antituberculeuses, l'infection par Mycobacterium tuberculosis est toujours l'un des principaux problèmes de santé publique au niveau mondial. Le traitement standard, bien qu'efficace, est long et difficile à mettre en œuvre. Par ailleurs, l'un des principaux challenges actuels est la prise en charge des formes résistantes de tuberculose, dont la prévalence est en augmentation constante. Nous avons souhaité mettre à profit les techniques récentes de modélisation in silico pour mieux comprendre l'action d'un des antituberculeux majeurs, l'isoniazide. L'objectif de ce travail est de construire un modèle mathématique global du traitement de la tuberculose pulmonaire par l'isoniazide basé sur des éléments pharmacocinétiques, pharmacodynamiques et physiopathologiques.En utilisant une approche de population, un modèle pharmacocinétique de diffusion pulmonaire a permis de décrire les concentrations en isoniazide dans le plasma et le poumon de 89 sujets. Des simulations réalisées à partir de ce modèle ont permis de montrer, par combinaison avec un modèle pharmacodynamique, qu'une individualisation de la posologie avec augmentation de la dose chez les métaboliseurs rapides de l'isoniazide n'aurait d'intérêt qu'en cas d'infection par une souche de M. tuberculosis de sensibilité intermédiaire. Ce modèle pharmacocinétique a ensuite été couplé à un modèle pharmacodynamique et à un modèle de la physiopathologie de la tuberculose pour construire un modèle mathématique thérapeutique global. Ce dernier a permis de simuler la dynamique bactérienne à partir du premier jour de l'infection et pour les deux premières semaines de traitement par isoniazide, et de reproduire certaines caractéristiques qualitatives et quantitatives de l'effet antibactérien initial de l'isoniazide connues en clinique. Les analyses réalisées sur ce modèle ont permis de mieux comprendre les déterminants de la dynamique bactérienne et notamment le ralentissement de l'effet antibactérien après les premiers jours de traitement / Despite the availability of a vaccine and several antituberculosis drugs, tuberculosis, an infection caused by Mycobacterium tuberculosis still remains a major public health concern worldwide. The standard treatment of tuberculosis is efficient but is long and hard to conduct. We decided to use recent in silico modeling techniques to improve our understanding of the action of isoniazid, one of the major antituberculosis drugs. The aim of this work was to build a full mathematical model reproducing the treatment of pulmonary tuberculosis by isoniazid, based on pharmacokinetics, pharmacodynamics and physiopathological data. A pulmonary pharmacokinetic model was developed in order to describe plasma and pulmonary isoniazid concentrations from 89 subjects. This model was used to perform simulations and was coupled to a pharmacodynamic model. The simulations showed that individualizing the dose according to the isoniazid metabolizer status would only be beneficial in fast metabolizers infected with a strain of intermediate sensitivity. This pharmacokinetic model was then integrated to a full mathematical model including a pharmacodynamic and an immune response model. This model adequately reproduced the bacterial dynamics over the development of the infection and its early treatment with isoniazid. It reproduced qualitative and quantitative features of the antimicrobial effect of isoniazid in agreement with clinical data. The simulations and analysis performed enabled us to better characterize the parameters influencing the bacterial dynamics and especially why the antibacterial effect tends to diminish after the first days of treatment

En dépit d’épidémies répétées, il n’existe pas à ce jour de thérapeutique ayant démontré son efficacité dans la maladie à virus Ebola. Sur la base d’expérimentations réalisées chez la souris et le macaque dans le cadre du consortium Reaction!, l’objectif de cette thèse visait à caractériser l’effet d’une molécule antivirale, le favipiravir, via l’implémentation de modèles mathématiques mécanistiques de l’infection et de la réponse immunitaire associée. L’approche utilisée pour construire ces modèles et en estimer les paramètres reposait sur les modèles non linéaires à effets mixtes. Un premier travail a permis d’explorer la relation concentration-effet sur la charge virale plasmatique chez la souris. Le second projet a conduit à caractériser la pharmacocinétique non linéaire dose et temps dépendante du favipiravir chez le macaque, en vue d’identifier les schémas posologiques pertinents pour la réalisation des études d’efficacité chez l’animal infecté. Au décours de leur réalisation, l’intégration des données virologiques et immunitaires générées au sein d’un modèle conjoint a permis de caractériser un effet modéré du favipiravir sur la réplication virale, mais suffisant pour limiter le développement d’une réaction inflammatoire délétère, et ainsi améliorer le taux de survie des animaux traités. Les simulations réalisées avec ce modèle ont pu souligner l’impact déterminant du délai d’initiation du traitement sur la survie. Ces résultats incitent à la poursuite de l’évaluation clinique du favipiravir, en favorisant des essais de prophylaxie ou post exposition. Enfin, un dernier travail a démontré l’absence de potentialisation du favipiravir par la ribavirine dans Ebola. / In spite of recurrent outbreaks, no therapeutics with demonstrated clinical efficacy are available in Ebola virus disease. Based on experimentations performed by Reaction! Consortium in mice and macaques, this thesis aimed to characterize the effect of an antiviral drug, favipiravir, using mechanistic mathematical models of the infection and associated immune response. The approach to build models and estimate parameters relied on nonlinear mixed effect models. The first project of this thesis explored the concentration-effect relationship on the viremia in mice. Then, a second project allowed to characterize the pharmacokinetics of favipiravir in macaques, underlying dose and time non linearity, and to identify relevant dosing regimen for efficacy experiments in infected animals. Once these experiments completed, the integration of the virological and immunological data into a mechanistic joint model shed light on the effect of favipiravir. The moderate inhibition of the viral replication resulting from the favipiravir plasma concentrations was enough to limit the development of a deleterious inflammatory response, and thus improve the survival rate of treated macaques. Simulations performed with this model underlined the crucial impact of the treatment initiation delay on survival. These results encourage the pursuit of the clinical evaluation of favipiravir in prophylaxis or post exposure trials. Finally, a last project demonstrated the lack of benefit of ribavirin addition to favipiravir in Ebola virus disease.